The
rodent
visual
system
has
attracted
great
interest
in
recent
years
due
to
its
experimental
tractability,
but
the
fundamental
mechanisms
used
by
mouse
represent
world
remain
unclear.
In
primate,
researchers
have
argued
from
both
behavioral
and
neural
evidence
that
a
key
step
representation
is
‘figure-ground
segmentation’,
delineation
of
figures
as
distinct
backgrounds.
To
determine
if
mice
also
show
signatures
figure-ground
segmentation,
we
trained
on
segmentation
task
where
were
defined
gratings
naturalistic
textures
moving
counterphase
background.
Unlike
primates,
severely
limited
their
ability
segment
figure
ground
using
opponent
motion
cue,
with
behavior
strongly
dependent
specific
carrier
pattern.
Remarkably,
when
forced
localize
patterns
motion,
they
adopted
strategy
brute
force
memorization
texture
patterns.
contrast,
including
humans,
macaques,
lemurs,
could
readily
independent
pattern
cue.
Consistent
behavior,
responses
same
stimuli
recorded
areas
V1,
RL,
LM
did
not
support
texture-invariant
motion.
Modeling
revealed
dependence
mouse’s
be
explained
feedforward
network
lacking
explicit
capabilities.
These
findings
reveal
limitation
objects
compared
primates.
Annual Review of Vision Science,
Год журнала:
2015,
Номер
1(1), С. 263 - 289
Опубликована: Ноя. 18, 2015
The
mammalian
retina
is
an
important
model
system
for
studying
neural
circuitry:
Its
role
in
sensation
clear,
its
cell
types
are
relatively
well
defined,
and
responses
to
natural
stimuli—light
patterns—can
be
studied
vitro.
To
solve
the
retina,
we
need
understand
how
circuits
presynaptic
output
neurons,
ganglion
cells,
divide
visual
scene
into
parallel
representations
assembled
interpreted
by
brain.
This
requires
identifying
component
interneurons
understanding
their
intrinsic
properties
synapses
generate
circuit
behaviors.
Because
cellular
composition
fundamental
of
shared
across
species,
basic
mechanisms
genetically
modifiable
mouse
apply
primate
vision.
We
propose
that
apparent
complexity
retinal
computation
derives
from
a
straightforward
mechanism—a
dynamic
balance
synaptic
excitation
inhibition
regulated
use-dependent
depression—applied
differentially
pathways
feed
cells.
Annual Review of Vision Science,
Год журнала:
2017,
Номер
3(1), С. 1 - 24
Опубликована: Июнь 15, 2017
Visual
signals
in
the
vertebrate
retina
are
shaped
by
feedback
and
feedforward
inhibition
two
synaptic
layers.
In
one,
horizontal
cells
establish
fundamental
center-surround
receptive-field
properties
via
morphologically
physiologically
complex
synapses
with
photoreceptors
bipolar
cells.
other,
a
panoply
of
amacrine
imbue
ganglion
cell
responses
spatiotemporally
information
about
visual
world.
Here,
I
review
current
ideas
signaling,
considering
evidence
for
against
leading,
competing
theories.
also
discuss
recent
work
that
has
begun
to
make
sense
remarkable
morphological
physiological
diversity
These
latter
efforts
have
been
aided
tremendously
increasingly
complete
connectivity
maps
inner
retinal
circuitry
new
genetic
tools
enable
study
individual,
sparsely
expressed
types.
Progress in Retinal and Eye Research,
Год журнала:
2020,
Номер
78, С. 100844 - 100844
Опубликована: Фев. 5, 2020
This
review
summarizes
our
current
knowledge
of
primate
including
human
retina
focusing
on
bipolar,
amacrine
and
ganglion
cells
their
connectivity.
We
have
two
main
motivations
in
writing.
Firstly,
recent
progress
non-invasive
imaging
methods
to
study
retinal
diseases
mean
that
better
understanding
the
is
becoming
an
important
goal
both
for
basic
clinical
sciences.
Secondly,
genetically
modified
mice
are
increasingly
used
as
animal
models
diseases.
Thus,
it
understand
which
extent
retinas
primates
rodents
comparable.
first
compare
cell
populations
rodent
retinas,
with
emphasis
how
fovea
(despite
its
small
size)
dominates
neural
landscape
retina.
next
summarise
what
known,
not
about
postreceptoral
neurone
The
inventories
bipolar
now
nearing
completion,
comprising
~12
types
at
least
17
cell.
Primate
show
clear
differences
dendritic
field
size
across
retina,
morphology
differs
clearly
from
mouse
cells.
Compared
cells,
even
higher
morphological
diversity:
they
could
comprise
over
40
types.
Many
appear
conserved
between
mice,
but
functions
only
a
few
understood
any
or
non-primate
Amacrine
final
frontier
research
monkeys
alike.
Annual Review of Vision Science,
Год журнала:
2018,
Номер
4(1), С. 165 - 192
Опубликована: Авг. 10, 2018
Visual
motion
on
the
retina
activates
a
cohort
of
retinal
ganglion
cells
(RGCs).
This
population
activity
encodes
multiple
streams
information
extracted
by
parallel
circuits.
Motion
processing
in
is
best
studied
direction-selective
circuit.
The
main
focus
this
review
neural
basis
direction
selectivity,
which
has
been
investigated
unprecedented
detail
using
state-of-the-art
functional,
connectomic,
and
modeling
methods.
Mechanisms
underlying
encoding
other
features
broader
RGC
populations
are
also
discussed.
Recent
discoveries
at
both
single-cell
levels
highlight
dynamic
stimulus-dependent
engagement
mechanisms
that
collectively
implement
robust
detection
under
diverse
visual
conditions.
Journal of Neuroscience,
Год журнала:
2016,
Номер
37(3), С. 610 - 625
Опубликована: Дек. 5, 2016
Retinal
ganglion
cells
(RGCs)
are
frequently
divided
into
functional
types
by
their
ability
to
extract
and
relay
specific
features
from
a
visual
scene,
such
as
the
capacity
discern
local
or
global
motion,
direction
of
stimulus
orientation,
contrast
uniformity,
presence
large
small
objects.
Here
we
introduce
three
previously
uncharacterized,
nondirection-selective
ON–OFF
RGC
that
represent
distinct
set
feature
detectors
in
mouse
retina.
The
high-definition
(HD)
RGCs
possess
receptive-field
centers
strong
surround
suppression.
They
respond
selectively
objects
sizes,
speeds,
motion.
We
present
comprehensive
morphological
characterization
HD
physiological
recordings
light
responses,
size
structure,
synaptic
mechanisms
also
explore
similarities
differences
between
well
characterized
with
comparably
receptive
field,
edge
detector,
response
moving
textures.
model
populations
each
type
study
how
they
differ
performance
tracking
object.
These
results,
besides
introducing
new
together
constitute
substantial
fraction
RGCs,
provide
insights
role
different
circuits
shaping
fields
establish
foundation
for
continued
suppression
neural
basis
motion
detection.
SIGNIFICANCE
STATEMENT
output
retina,
retinal
(RGCs),
diverse
group
∼40
neuron
often
assigned
“feature
detection”
profiles
based
on
aspects
scene
which
respond.
describe,
first
time,
substantially
augmenting
our
understanding
selectivity.
Experiments
modeling
show
while
these
“high-definition”
share
certain
properties,
have
tuning
size,
speed,
enabling
them
occupy
niches
space.